Kinetic and Thermodynamic Acidities of Pentacarbonyl(cyclobutenylidene)chromium Complexes. Effect of Antiaromaticity in the Conjugate Anion. An Experimental and Computational Study† Claude F. Bernasconi,*,‡ Viola Ruddat,‡ Philip J. Wenzel,‡ and Helmut Fischer§ Department of Chemistry and Biochemistry of the University of California, Santa Cruz, California 95064, and the Department of Chemistry of the University of Konstanz, D-78434 Konstanz, Germany
[email protected] Received April 5, 2004
The deprotonation of pentacarbonyl[(3-diethylamino-2,4-dimethyl)cyclobut-2-ene-1-ylidene]chromium (1d) and pentacarbonyl[(3-diethylamino-4-methyl-2-phenyl)cyclobut-2-ene-1-ylidene]chromium (1e) leads to antiaromatic conjugate anions by virtue of their being cyclobutadiene derivatives. Rate constants for the deprotonation of 1d and 1e by P2-Et and pKa values were determined in acetonitrile. Gas-phase B3LYP calculations of 1d, 1e, and their respective conjugate anions, using a generalized basis set, were also performed. Furthermore, for purposes of comparison with carbene complexes of similar structures, but having conjugate anions that are not antiaromatic, corresponding calculations were performed on pentacarbonyl[3-diethylamino-2,5-dimethyl)cyclopent-2ene-1-ylidene]chromium (5), [dimethylamino(methyl)carbene]pentacarbonylchromium (3a), and [dimethylamino(iso-propyl)carbene]pentacarbonylchromium (3b) and their respective conjugate anions, and solution-phase pKa and kinetic measurements were carried out for 3a and 3b. Major points of interest include the effect of antiaromaticity on the kinetic and thermodynamic acidities of 1d and 1e, the large effect of the phenyl group on the gas-phase acidity of 1e, the strong attenuation of the acidities and the effect of the phenyl group in acetonitrile, and the position of the CdC double bonds in the cyclobutadiene ring of the conjugate anion of 1e. Introduction The synthesis and detailed characterization of the pentacarbonyl(cyclobutenylidene)chromium complexes
these Fischer-type carbene complexes2 derives from their nonlinear optical properties which are related to the push-pull electronic structure symbolized by the strong contribution of the resonance form B.1
Two of these complexes, 1d and 1e, have an acidic proton which can be removed in the presence of a strong
1a-e has recently been reported.1 A primary interest in † Physical Organic Chemistry of Transition Metal Carbene Complexes. 30. Part 29: Bernasconi, C. F.; Bhattacharya, S. Organometallics 2004, 23, 1722. ‡ University of California. § University of Konstanz. (1) Fischer, H.; Podschadly, O.; Roth, G.; Herminghaus, S.; Klewitz, S.; Heck, J.; Houbrechts, S.; Meyer, T. J. Organomet. Chem. 1997, 541, 321.
(2) Do¨tz, K. H.; Fischer, H.; Hofmann, P.; Kreissl, F. R.; Schubert, U.; Weiss, K. Transition Metal Complexes; Verlag Chemie: Deerfield Beach, FL, 1983. (3) (a) Bernasconi, C. F.; Chem. Soc. Rev. 1997, 26, 299. (b) Bernasconi, C. F. Adv. Phys. Org. Chem. 2002, 37, 137. (4) As shown below, ab initio calculations on 1e- indicate double bonds C1-C3 and C2-C4 rather than C1-C2 and C3-C4; for 1dwhich is symmetrical both representations are equivalent. (5) The intrinsic barrier, ∆Goq (intrinsic rate constant, ko), of a reaction with a forward rate constant k1 and a reverse rate constant q k-1 is defined as ∆Goq ) ∆G1q ) ∆G-1 (ko ) k1 ) k-1) when ∆Go ) 0 (K1 ) 1). (6) Bernasconi, C. F.; Leyes, A. E.; Ragains, M. L.; Shi, Y.; Wang, H.; Wulff, W. D. J. Am. Chem. Soc. 1998, 120, 8632. 10.1021/jo040168t CCC: $27.50 © 2004 American Chemical Society
5232
J. Org. Chem. 2004, 69, 5232-5239
Published on Web 07/14/2004
Acidities of Pentacarbonyl(cyclobutenylidene)chromium Complexes
base, eq 1. What is unique about these systems is that,
TABLE 1. pKa Values and Rate Constants for the Deprotonation of 1d, 1e, 3a, and 3b by P2-Et in Acetonitrile 25 °C carbene complex
pKa
1d 1e 3a 3b 2 (R′ ) R′′ ) H)b
34.6 ( 0.2 33.9 ( 0.2 31.9 ( 0.4 32.4 ( 0.3 22.2 ( 0.2
a pKcorr a
k1 (M-1 s-1) kcorr (M-1 s-1)a 1
34.3 ( 0.2 10.0 ( 0.5 10.0 ( 0.5 33.9 ( 0.2 7.24 ( 0.25 7.24 ( 0.25 32.4 ( 0.4 210 ( 20 70 ( 7 32.4 ( 0.3 0.183 ( 0.013 0.183 ( 0.013 22.7 ( 0.2
a Statistically corrected for the number of equivalent protons of the carbene complex and equivalent basic sites on the anion. b Reference 6.
unlike other Fischer carbene complexes that contain acidic protons such as 2 and similar compounds,3 deprotonation of 1d and 1e generates conjugate bases (1d-, 1e-) that are antiaromatic by virtue of their cyclobutadiene ring.4 Hence, one would expect that the thermodynamic acidity of 1d and 1e should be depressed relative to that of other carbene complexes.
What is less predictable is the effect of the antiaromaticity on the intrinsic barrier or intrinsic rate constant5 of the reaction, i.e., the kinetic acidity. In this paper, we present experimental data on the kinetic and thermodynamic acidities of 1d and 1e in acetonitrile and results of gas phase ab initio calculations of the structures and thermodynamic acidities of these carbene complexes. We show that in the gas phase the effect of antiaromaticity on the thermodynamic acidities is large but strongly attenuated in acetonitrile. We also conclude that the intrinsic barrier to proton transfer is relatively low, which suggests that antiaromaticity may have a barrier-lowering effect. Results Spectrophotometric pKa Determination. 1d and 1e are much weaker acids than typical Fischer carbene complexes such as 36 but they can be deprotonated by 1-ethyl-2,2,4,4,4-pentakis(dimethylamino)-2R,54R5-catenadi(phosphazene), P2-Et, whose pKa in acetonitrile is 32.8.7
lowed a spectrophotometric determination of their pKa values as described in the Experimental Section. In contrast to the situation with 2 and similar carbene complexes, the deprotonation of 1d and 1e is relatively slow. Because the respective anions are not very stable, this means that there is some decomposition of the anions before the deprotonation is complete, necessitating extrapolation of the absorbance values to “zero time.” For comparison purposes, the acidity of 3b was also determined, using the same methodology; the pKa of 3a has been determined before.8 The various pKa values are reported in Table 1.
Rates of Deprotonation. Kinetic experiments were run under pseudo-first-order conditions with a large excess of P2-Et. The reaction was shown to be strictly first order in P2-Et over a concentration range from 4 × 10-4 to 3 × 10-3 M. Second-order rate constants, k1, for the deprotonation reaction were obtained from the slopes of kobsd versus [P2-Et]; they are reported in Table 1. Calculations. DFT calculations using the B3LYP density functional were performed on 1d, 1e, 3a, and 3b and their respective anions 1d-, 1e-, 3a-, and 3b-. In our experimental study we have used 3a and 3b as models for carbene complexes with a strong π-donor group (Me2N) but whose respective conjugate bases are not antiaromatic and hence may be compared with 1d and 1e to assess the effect of antiaromaticity on acidity and rate constants. This choice of models is not optimal but was dictated by the availability/accessibility of known Fischer carbenes. A better model is the carbene complex 5 which is both cyclic and includes the same push-pull system as 1d and 1e. Even though 5 is not a known compound, calculations on it and its conjugate base 5were performed.
The deprotonation of 1d and 1e is associated with a decrease in absorption and a hypsochromic shift. This alJ. Org. Chem, Vol. 69, No. 16, 2004 5233
Bernasconi et al.
FIGURE 1. Structures of 1d, 1d-, 1e, and 1e-. FIGURE 3. Structures of 3a, 3a-, 3b, and 3b-.
the optimized structure with the double bonds C1-C3 and C2-C4, and 1e-*, a higher energy structure, with double bonds constrained to C1-C2 and C3-C4. Note that for the anion of 1d the two structures, 1d- and 1d-*, are identical. Table 2 also includes experimental bond lengths for the model system 1b.
FIGURE 2. Structures of 5 and 5-.
To keep the calculations at a manageable level for 1d, 1d-, 1e, 1e-, 5, and 5-, a generalized basis set was used as described in the Experimental Section. For consistency, the same generalized basis set was also used for 3a, 3a-, 3b, and 3b-. Figure 1 shows the optimized (at B3LYP level) structures of 1d, 1d-, 1e, and 1e-, respectively, while the structures of 5 and 5- are displayed in Figure 2 and those of 3a, 3a-, 3b and 3b- in Figure 3. Tables 2-4 summarize relevant bond lengths and dihedral angles. For the anion of 1e two structures were calculated: 1e- is (7) Dega-Szafran, Z.; Schroeder, G.; Szafran, M. J. Phys. Org. Chem. 1999, 12, 39. (8) Bernasconi, C. F.; Ragains, M. L. To be published.
5234 J. Org. Chem., Vol. 69, No. 16, 2004
NPA group charges (at the B3LYP level) are summarized in Figures 4-6 while Table 5 reports the gasphase acidities. More complete computational results are summarized in Tables S1-S3 (geometries) and Table S4 (energies) of the Supporting Information.9 Discussion Computational Results. The computed bond lengths for 1d and 1e are quite similar to each other as well as to those calculated and determined by X-ray crystal(9) See the Supporting Information paragraph.
Acidities of Pentacarbonyl(cyclobutenylidene)chromium Complexes TABLE 2. Bond Lengths and Dihedral Angles (B3LYP) for 1d, 1d-, 1e, 1e-, 1e-*, and 1ba bond or dihedral angle
1d
1d-
CdO trans CdO cis Cr-CO trans Cr-CO cis CrdC1 C1-C3 C1-C2 C3-C4 C2-C4 C4-N C3-C5 CO cis 1-Cr-C1-C2 CO cis 2-Cr-C1-C2 Cl-C3-C5-C6 C1-C2-C3-C4
1.173 1.175 1.886 1.882 2.003 1.406 1.557 1.412 1.517 1.329
1.183 1.178 1.836 1.867 2.091 1.352 1.588 1.563 1.354 1.389
a
-42.4 48.4
-39.0 50.8
-4.5
-0.7
Bond lengths in angstroms, angles in degrees.
b
1d-
∆b
1e
1e-
∆c
1e-*
1b (exp)d
0.010 0.003 -0.050 -0.015 0.088 -0.054 0.031 0.151 -0.163 0.060
1.173 1.175 1.888 1.884 1.994 1.408 1.558 1.412 1.520 1.326 1.473 -80.0 10.5 73.5 10.1
1.183 1.180 1.853 1.867 2.076 1.375 1.566 1.563 1.352 1.389 1.435 -34.8 55.3 142.7 -0.4
0.010 0.005 -0.035 -0.017 0.082 -0.033 0.008 0.151 -0.168 0.063 -0.038
1.184 1.180 1.850 1.875 2.085 1.587 1.365 1.381 1.523 1.364 1.429 -39.41 51.82 142.1 4.12
n.r. n.r. 1.815 1.902 2.100 1.390 1.572 1.428 1.542 1.315 1.512 -97.7 10.3
- 1d.
c
TABLE 3. Bond Lengths and Dihedral Angles (B3LYP) for 5 and 5- a
bond or dihedral angle CdO trans CdO cis Cr-CO trans Cr-CO cis
CrdC1 C1-C3 C1-C2 C3-C4 C4-C5 C2-C5 C4-N CO cis 1-Cr-C1-C3 CO cis 2-Cr-C1-C3 C1-C3-C4-C5 C2-C1-C3-C4 a
5
5-
∆b
1.175 1.176 1.882 1.889 1.802 1.910 1.889 2.095 1.409 1.527 1.420 1.504 1.541 1.341 -28.9 63.2 -4.5 5.6
1.186 1.181 1.852 1.878 1.871 1.875 1.897 2.201 1.504 1.367 1.361 1.501 1.511 1.433 -53.2 38.5 -1.2 0.8
0.011 0.005 -0.030 -0.011 -0.011 -0.035 -0.008 0.106 0.095 -0.160 -0.059 -0.003 -0.030 0.092
Bond lengths in angstroms, angles in degrees. b 5- - 5.
lography for 1b.1 In particular, the C1-C3 and C3-C4
bonds are much shorter than the C1-C2 and C2-C4 bonds due to their partial double-bond character. As pointed out before,1 the C1-C3 bond is even somewhat shorter than the C3-C4 bond, implying that resonance structure B is dominant. The respective bond lengths for 5 are also quite similar to those for 1d although the CrC1, C1-C3, C3-C4, and C4-N bonds in 5 are all slightly longer than the corresponding bonds in 1d while the C1C2 bond is somewhat shorter. A detailed interpretation of these differences is difficult to come by except to say that they must be related to the change from a fourmembered to a five-membered ring. The changes in the bond lengths upon conversion of the cyclobutenylidene complexes into their respective anions are also quite similar for 1d and 1e. The lengthening of the CdO and shortening of the Cr-CO bond lengths are the result of charge delocalization into the
1e-
- 1e.
d
6.7
From X-crystal structure data, ref 1.
CO ligands, while the lengthening of the CrdC1 and the C4-N bonds reflects the loss of their partial double bond character. The bond changes within the four-membered ring are particularly noteworthy because the shortening of the C1-C3 and C2-C4 implies that the position of the double bonds in the anion of 1e is that in 1e- rather than in 1e-*; as discussed below, 1e- is 0.78 kcal/mol lower in energy than 1e-*. For the conversion of 5 to 5-, similar changes in the CdO, Cr-CO, Cr-C1, and C4-N bonds are observed as for the conversion of 1d to 1d- and 1e to 1e-. In contrast, the C1-C3 bond of 5- is lengthened while the C1-C2 and C3-C4 bonds are shortened, reflecting the facts that in 5- it is the C1-C2 and C3-C4 bonds rather than the C1-C3 and C2-C4 bonds that adopt double bond character. For 3a and 3b, the pattern in the bond changes upon conversion to the respective anions are as one would expect: slight lengthening of the CdO bonds, shortening of the Cr-CO bonds, substantial lengthening of the CrC1 and C1-N bonds and substantial shortening of the C1-C2 bonds. Turning to the gas-phase acidities reported in Table 5, we note that 1d and 1e are substantially less acidic than 5, 3a and 3b. The largest difference is between 1d and 5 with 1d being 35 kcal/mol less acidic than 5. The low acidity of 1d and 1e can be attributed to the destabilization of 1d- and 1e- due to their antiaromatic character.10 The higher acidity of 1e compared to that of 1d (∆∆H° ) 7.8 kcal/mol) suggests that the phenyl group in 1eacts as a π-acceptor (1e- T 1e-′). This interpretation is
supported by three observations. (1) The C1-C3 bond (10) According to a recent MP4(SDQ)/6-31G(d,p) calculation the absolute antiaromaticity of cyclobutadiene is 40.28 kcal/mol.11
J. Org. Chem, Vol. 69, No. 16, 2004 5235
Bernasconi et al. TABLE 4. Bond Lengths and Dihedral Angles (B3LYP) for 3a, 3a-, 3b, and 3bbond or dihedral angle
3a
3a-
∆b
3b-
3b
∆c
3a (exp)
CdO trans 1.173 1.183 0.010 1.173 1.816 0.013 1.160 CdO cis 1.174 1.185 0.011 1.175 1.186 0.011 1.130 Cr-CO trans 1.880 1.858 -0.022 1.876 1.844 -0.032 1.850 Cr-CO cis 1.883 1.843 -0.040 1.885 1.883 -0.002 1.860 Cr-C1 2.084 2.189 0.105 2.127 2.195 0.068 2.160 C1-C2e 1.520 1.356 -0.164 1.330 1.351 -0.184 1.500 C1-N 1.327 1.429 0.102 1.535 1.450 0.120 1.310 CO cis 145.41 -9.03 -21.88 1.00 Cr-C1-C2 CO cis 246.54 80.94 68.61 88.4 Cr-C1-C2 a
Bond lengths in Å, angles in degrees.
b
3a-
- 3a.
c
3b-
- 3b.
FIGURE 5. NPA group charges (at B3LYP level) for 5 and 5-.
TABLE 5. Gas Phase Acidities at 25°C ∆E (kcal/mol(
∆ZPE (kcal/mol)
∆H° (kcal/mol)
1d f 1d- + H+
378.6
-9.0
369.6
H+
1e f + 3a f 3a- + H+
370.8 348.6
-9.0 -7.7
361.8 340.9
3b f 3b- + H+ 5 f 5- + H+
345.9 344.5
-9.9 -9.7
336.0 334.8
1e-
a
∆∆H° (kcal/mol) 7.8a 4.9b
∆H°(1d) - ∆H°(1e). b ∆H°(3a) - ∆H°(3b).
FIGURE 6. NPA group charges (at B3LYP level) for 3a, 3a-, 3b, and 3b-.
FIGURE 4. NPA group charges (at B3LYP level) for 1d, 1d-, 1e, 1e-, and 1e-*.
contraction upon deprotonation of 1e (-0.033 Å) is smaller than that for the deprotonation of 1d (-0.054 Å). (2) The deprotonation of 1e induces the phenyl group to change its orientation from nearly perpendicular (dihedral angle C1-C3-C5-C6 of 73.5°) to more coplanar (dihedral angle C1-C3-C5-C6 of 142.7°). (3) As is shown in Figure 4, there is a significant negative charge on the phenyl group of 1e- (-0.180) while the negative charge on the (CO)5Cr moiety (-0.587) is significantly lower than that on 1d- (-0.645). The fact that in the conjugate base of 1e the preferred position of the double bonds in the cyclobutadiene ring is that of structure 1e- rather than 1e-* suggests a mechanism of the deprotonation reaction. Two mecha5236 J. Org. Chem., Vol. 69, No. 16, 2004
nistic pathways may be visualized, eqs 3 and 4, depending on which resonance structure of 1d or 1e is chosen to represent the carbene complex. Our results support the mechanism of eq 4.12
Acidities of Pentacarbonyl(cyclobutenylidene)chromium Complexes
Regarding possible reasons why 1e- is more stable than 1e-*, we note the following. As pointed out above, charge delocalization into the phenyl ring (1e-′) helps stabilize 1e- but the NPA group charges indicate a similar delocalization of negative charge into the phenyl group of 1e-* that may be understood in terms of the resonance structure 1e-*′. The fact that the C4-N bond
in 1e-* (1.364 Å) is shorter than 1e- (1.389 Å) is consistent with a contribution by 1e-*′. Hence, the main reason for the different stabilities is probably not electronic but steric, i.e., steric crowding between the phenyl group and the Et2N group may be more severe than that between the phenyl group and the (CO)5Cr moiety. Thus, elongating the C3-C4 rather than the C1-C3 bond would bring about greater relief of steric strain. Regarding 3a and 3b, the structures of these carbene complexes are too different from that of 5 to allow a detailed explanation as to why their acidities are somewhat lower than that of 5. However, a comparison between 3a and 3b is in order. The higher acidity of 3b relative to that of 3a (∆∆H° ) 4.9 kcal/mol) is probably the combined result of a destabilization of the zwitterionic structure 3b′ due to steric interference by the isopropyl group and an enhanced stabilization of the CdC double bond in 3b- by the methyl groups.14 A reduced contribution of 3b′ to the resonance hybrid is consistent with the smaller negative charge on the (CO)5Cr moiety and the smaller positive charge in the Me2N group of 3b relative to the same charges in 3a.
(31.9) (Table 1) which, in enthalpy terms, corresponds to a 3.5 kcal/mol difference. The main difference between solution and gas phase must be the solvation of the charges in the partially zwitterionic carbene complexes and their respective anions. The solvation of 1d- and 3a- is probably quite similar. On the other hand, solvational stabilization of 3a is expected to be stronger than that of 1d because the more extended π-system in 1d provides more internal stabilization. This results in a reduction of the acidity of 3a relative to that of 1d. The acidity difference between 1d and 1e, and that between 3a and 3d, are also strongly attenuated in acetonitrile. For 1d versus 1e this attenuation may be related to a stronger solvation of 1d- because in 1d- the negative charge is more concentrated on the (CO)5Cr moiety than in 1e- (Figure 4), and hence, the strong solvation partially offsets the greater internal stabilization of 1e- by the phenyl group that manifests itself in the gas phase. With 3a versus 3b, the steric destabilization of 3b′ that occurs in the gas phase may be attenuated because solvation can support larger charges on the (CO)5Cr and Me2N moieties, resulting in a lower acidity of 3b and making the pKa values of 3a and 3b about the same. A question of major interest is whether the fact that the deprotonation of 1d and 1e leads to anions with an antiaromatic cyclobutadiene ring has a significant effect on the intrinsic barriers or intrinsic rate constants5 of these reactions, and whether a comparison between 1d and 1e on one hand and 3a and 3b on the other allows such a determination. If the development of antiaromaticity at the transition state has made less progress than proton transfer, there should be a decrease in the intrinsic barrier or an increase in the intrinsic rate constant; if the development of antiaromaticity has made more progress than proton transfer at the transition state, ∆Goq should increase and ko should decrease.16 It has recently been suggested that development of aromaticity is ahead of proton transfer in the deprotonation of cyclic rhenium carbene complexes such as 4H+(X) where the conjugate bases 4(X) contain the furan, thiophene and selenophene rings, respectively.18 This
Results in Acetonitrile. In acetonitrile the acidity differences between the cyclobutenylidene complexes (1d, 1e) and dimethylamino carbene complexes (3a, 3b) are much smaller than in the gas phase; e.g., the measured15 pKa of 1d (34.6) is only 2.7 units higher than that of 3a
conclusion was somewhat surprising because it is op-
(11) Suresh, C. H.; Koga, N. J. Org. Chem. 2002, 67, 1965. (12) A reviewer suggested that eqs 3 and 4 may not be really different because they are related via resonance. It is our contention that they are different because 1e- and 1e-* are not resonance structures but different compounds, a well-known feature of cyclobutadienes.13 (13) Minkin, V. I.; Glukhovtsev, M. N.; Simkin, B. Y. Aromaticity and Antiaromaticity; Wiley & Sons: New York, 1994; pp 138-141. (14) The stabilization of alkenes by methyl groups is a well-known phenomenon, see, e.g.: McMurry, J. Organic Chemistry; Brooks/Cole: Belmont, CA, 2003; p 183. (15) When calculating and comparing enthalpies, the use of the experimental, i.e., statistically uncorrected pKa values, is more appropriate.
(16) These predictions are based on the Principle of Nonperfect Synchronization (PNS).17 The PNS states that the intrinsic barrier of a reaction increases if the development of a product stabilization factor (e.g., resonance or solvation) lags behind bond changes but decreases if the development of the product stabilization factor runs ahead of bond changes. Conversely, if a product destabilizing (e.g., antiaromaticity) lags behind bond changes the intrinsic barrier decreases while the intrinsic barrier increases if the product destabilizing factor is ahead of bond change. (17) (a) Bernasconi, C. F. Acc. Chem. Res. 1987, 20, 301. (b) Bernasconi, C. F. Acc. Chem. Res. 1992, 25, 9. (c) Bernasconi, C. F. Adv. Phys. Org. Chem. 1992, 27, 119. (18) Bernasconi, C. F.; Ragains, M. L.; Bhattacharya, S. J. Am. Chem. Soc. 2003, 125, 12328.
J. Org. Chem, Vol. 69, No. 16, 2004 5237
Bernasconi et al.
posite to the situation with ordinary resonance/delocalization effects, i.e., the development of these latter effects always lags behind proton transfer at the transition state.17,19 Furthermore, ab initio calculations of gas-phase identity proton transfers in systems such as benzenium ion/benzene and cyclopentadiene/cyclopentadienyl anion also indicated that aromaticity is more than 50% developed at the transition state.23 These results suggest that the constraints that prevent extensive development of “ordinary” resonance stabilization at the transition state17,21d,24 may not apply to the development of aromaticity. Because of the interrelatedness of aromaticity and antiaromaticity it would seem plausible that the latter would follow the pattern observed with the former, i.e., antiaromaticity develops ahead of proton transfer, thereby raising the intrinsic barrier or lowering the intrinsic rate constant. On the other hand, if the development of antiaromaticity were to lag behind proton transfer, as is the case for ordinary resonance effects, this could lower the energy of the transition state and hence would be a more advantageous situation. Our results do not allow a definite conclusion on this point because several factors other than antiaromaticity may affect the intrinsic rate constants of the reactions of 1d and 1e relative to those of the reactions of 3a and 3b. Nevertheless, the following discussion should provide value for the deprosome important insights. The kcorr 1 tonation of 1d by P2-Et is 7-fold lower than for 3a (Table 1) but this reduction can be accounted for by the higher 25 pKcorr a . Applying the simplest version of the Marcus equation to the ratio of the rate constants, eq 6, yields
log
ko (1d) ko (3a)
)
kcorr 1 (1d) kcorr 1 (3a)
+ 0.5 log
kcorr a (3a) kcorr a (1d)
(6)
corr log(kcorr o (1d)/ko (3a) ) 0.105, i.e., the intrinsic rate constants for the two reactions are virtually the same. An interpretation of this result requires an examination of the factors that distinguish 1d and 1e from 3a, factors that may potentially affect the relative intrinsic rate constants for proton transfer. These factors include the overall structure of the respective carbene complexes, steric effects, the degree of push-pull resonance and the antiaromaticity of 1d- and 1e-. 1. Overall Structure. The structures of 1d and 1e are sufficiently different from that of 3a that they may lead to different “baseline” intrinsic rate constants for 1d and 1e relative to 3a. If this is the case we cannot predict which “baseline” intrinsic barrier may be the higher one.
(19) This is also true for gas-phase proton transfers.20-22 (20) (a) Saunders, W. H., Jr. J. Am. Chem. Soc. 1994, 116, 5400. (b) Saunders, W. H., Jr.; Van Verth, J. E. J. Org. Chem. 1995, 60, 3452. (21) (a) Bernasconi, C. F.; Wenzel, P. J. J. Am. Chem. Soc. 1994, 116, 5405. (b) Bernasconi, C. F.; Wenzel, P. J. J. Am. Chem. Soc. 1996, 118, 10494. (c) Bernasconi, C. F.; Wenzel, P. J.; Keeffe, J. R.; Gronert, S. J. Am. Chem. Soc. 1997, 119, 4008. (d) Bernasconi, C. F.; Wenzel, P. J. J. Org. Chem. 2001, 66, 968. (22) Lee, I.; Kim, C. K.; Kim, C. K. J. Phys. Org. Chem. 1999, 12, 255. (23) Bernasconi, C. F.; Ragains, M. L. To be published. (24) Kresge, A. J. Can. J. Chem. 1974, 52, 1897. (25) (a) Marcus, R. A. Annu. Rev. Phys. Chem. 1964, 15, 155. (b) Marcus, R. A. J. Phys. Chem. 1968, 72, 891.
5238 J. Org. Chem., Vol. 69, No. 16, 2004
2. Steric Effect. The transition state for the deprotonation of 1d and 1e is expected to be more crowded than that for the deprotonation of 3a because the proton in 1d and 1e is on a tertiary carbon. This will reduce ko for 1d and 1e relative to that for 3a. Due to the large size of the base (P2-Et), this reduction may be substanfor the more crowded tial as the large decrease in kcorr 1 3b relative to 3a suggests. 3. Push-Pull Resonance. The effect of this factor on ko is difficult to predict because two opposing influences are involved. One is the loss of the resonance stabilization of the carbene complex (A T B) upon deprotonation which must follow the general rule that applies to resonance effects, i.e., its loss must be ahead of proton transfer at the transition state.17,19 The result is a reduction in ko. The other effect of the push-pull resonance of the carbene complex is an attenuation of the lag in resonance development of the anion. This comes about because the contribution of B (for 1d and 1e) or 3′ to the structure of the respective carbene complexes leads to a preorganization of the (CO)5Cr moiety toward its electronic configuration in the anion. This should result in a stabilization of the transition state due to greater charge delocalization into the (CO) 5Cr moiety, leading to a lower intrinsic barrier or a larger ko. As discussed in more detail elsewhere,26 there are precedents for either of these factors to be dominant. It appears that in systems with the strongest push-pull resonance, the second factor is dominant which leads to an increase in ko,27 while in systems where the push-pull resonance is weaker, the first factor dominates and ko is reduced.28 To the extent that the push-pull effect is particularly strong in 1d and 1e one might expect ko for 1d and 1e to increase relative to that for 3a but this conclusion is not firm. 4. Antiaromaticity. Based on our earlier discussion, ko for 1d and 1e should be enhanced relative to ko for 3a if antiaromaticity lags behind proton transfer but should be reduced if antiaromaticity develops ahead of proton transfer. The fact that the experimental ko values for 1d and 3a are about the same despite the steric effect implies that one, two or all three factors discussed above, i.e., overall structure, push-pull resonance and antiaromaticity act to offset the steric effect by enhancing ko for 1d and 1e. Unless the “baseline” ko values for 1d and 1e are significantly higher than for 3a, it is unlikely that the push-pull resonance effect alone is sufficient to offset the steric effect. This would suggest that the antiaromaticity factor contributes to the enhancement of ko, implying that antiaromaticity development lags behind proton transfer. However, further work on systems where factors other than antiaromaticity can be excluded will be needed to find a more definite answer. Conclusions (1) In the gas phase 1d and 1e are much less acidic than 5, 3a, or 3b, presumably mainly because 1d- and (26) Bernasconi, C. F.; Ali, M. J. Am. Chem. Soc. 1999, 121, 3039. (27) (a) Bernasconi, C. F.; Renfrow, R. A.; Tia, P. R. J. Am. Chem. Soc. 1986, 108, 4541. (b) Bernasconi, C. F.; Zitomer, J. L.; Schuck, D. F. J. Org. Chem. 1992, 57, 1132. (28) (a) Bernasconi, C. F.; Panda, M. J. Org. Chem. 1987, 52, 3042. (b) Bernasconi, C. F.; Killion, R. B., Jr. J. Org. Chem. 1989, 54, 2878. (c) Bernasconi, C. F.; Flores, F. X.; Claus, J. J.; Dvor`a´k, J. J. Org. Chem. 1994, 59, 4917.
Acidities of Pentacarbonyl(cyclobutenylidene)chromium Complexes
1e- are antiaromatic. Solvation effects in acetonitrile strongly attenuate these acidity differences. (2) Replacing the methyl group in 1d with a phenyl group (1e) enhances the gas-phase acidity by 7.8 kcal/ mol. This can be attributed to partial delocalization of the anionic charge into the phenyl group, consistent with computed bond lengths, dihedral angles, and NPA group charges. (3) The anionic structure of 1e- is 0.78 kcal/mol more stable than 1e-*. A possible reason is that a longer C3C4 bond (1e-) alleviates more steric strain than a longer C1-C3 bond (1e-*). (4) The measured rate constants for the deprotonation of 1d by P2-Et suggest that the intrinsic rate constant is about the same as for the deprotonation of 3a by the same base. However, in the absence of steric effects, the “true” ko value for 1d would be significantly higher than for 3a. A possible interpretation of this enhanced intrinsic rate constant is that the development of antiaromaticity lags behind proton transfer at the transition state. Because of other possible contributions such as the “baseline” ko value and the push-pull effect, this interpretation can only be tentative.
Experimental Section Materials. The pentacarbonyl(cyclobutenylidene)chromium complexes 1d and 1e were available from a previous study.1 3b was synthesized as described in the literature.29 The product was dried and purified by column chromatography with 90:10 hexane ether. The solution volume was reduced, crystals were collected at 0°C and dried under vacuum. 1H NMR δ (CDCl3): 1.23 (6H, d, (CH3) 2), 3.45 (1H, m, H), 3.95 (6H, 2s, N(CH3) 2). P2-Et was purchased from Fluka. Acetonitrile was dried over P2O5 and distilled under nitrogen. pKa Measurements. A detailed description of the methodology has been reported before,6 and hence, only a brief summary is presented here. The pKa of the carbene complexes + (pKCH a ) is given by eq 7 where BH refers to the protonated form of P2-Et and K1, the equilibrium constant for the ionization equilibrium CH + B a C- + BH+, is defined by eq 8. BH+ pKCH - log K1 a ) pKa
K1 )
[C-] + [BH+] [CH][B]
(7) (8)
The concentration of the free carbene complex ([CH]) is related to its total concentration ([CH]o) by eq 9 where r is given by eq 10 and A refers to the absorbance at intermediate P2-Et
[CH] ) r[CH]o r)
A - A∞ A o - A∞
(9) (10)
concentrations (partial conversion of CH to its anion C-), Ao to the absorbance in the absence of P2-Et (no conversion of CH to C-), and A∞ to the absorbance of high [P2-Et] (complete conversion of CH to C-). The concentration of the other species (29) (a) Aumann, R.; Fischer, E. O. Angew. Chem., Int. Ed. Engl. 1967, 6, 879. (b) Connor, J. A.; Fischer, E. O. J. Chem. Soc. A 1969, 578.
are calculated as shown in eqs 11 and 12.
[C-] ) [BH+] ) (1 - r)[CH]o
(11)
[B] ) [B]o - [BH+]
(12)
K1 values for a given carbene complex were determined for 11-16 different sets of [CH]o and [B]o; the raw data are summarized in Tables S5-S7 of the Supporting Information.9 The pKCH values reported in Table 1 represent the averages a of these determinations. Computational Studies. All computations were carried out using the GAUSSIAN 98 suite of programs.30 Calculations were performed at the HF and B3LYP levels of theory using a generalized basis set. The Way-Hadt effective core potential31 was assigned to the chromium atom, the 6-31+G* basis set was used for the atoms that are part of the reacting system (1d, 1d-, 1e, and 1e-: four-membered ring atoms, N; 3a and 3b: C of CHR2, N; 5 and 5-: five-memberered ring atoms, N), while the 3-21G* basis set was assigned to the atoms that are not part of the reacting system (1d and 1e: Me, R′′, Et2, H, (CO) 5; 1d- and 1e-: Me, R′′, Et2, (CO)5; 3a and 3b: Me2, (CO)5, H; 3a- and 3b-: Me2, (CO)5; 5 and 5-: Me, Et2, (CO)5). The geometries of all neutral carbene complexes and their respective anions were first optimized at the HF level and then further optimized at the B3LYP level. For 1e-*, the initial geometry was constrained in such a way that the length of the C1-C2 bond corresponded to that of the B3LYP C1-C3 bond in 1e-, the C1-C3 bond to that of the B3LYP C1-C2 bond in 1e-, and the C2-C4 to that of the B3LYP C3-C4 bond in 1e-. The structure of 1e-* constrained in this manner was then optimized at the B3LYP level. Frequency calculations were performed at the B3LYP level and contributions to the vibrational partition functions were scaled with the scaling factor 0.9613 according to Scott and Radom.32 The acidities were calculated as
∆H° ) ∆E + ∆ZPE + ∆CvibRT
(13)
where ∆E is the difference in electronic energy at a given level of theory, ∆ZPE the difference in the scaled zero-point energies and ∆CvibRT the difference in the scaled vibrational heat capacities.
Acknowledgment. This research has been supported by Grant No. CHE-0098553 from the National Science Foundation. Partial support of V.R. by the Department of Education in the form of a GAANN fellowship is also gratefully acknowledged. Supporting Information Available: Tables S1-S3 (geometries), S4 (energies), and S5-S7 (pKa determinations) and structures S1-S10 (Z-matrixes). This material is available free of charge via the Internet at http://pubs.acs.org. JO040168T (30) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann, R. E., Jr.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. GAUSSIAN 98, Revision a.7; Gaussian, Inc., Pittsburgh, PA, 1998. (31) (a) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270. (b) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270. (32) Scott, A.; Radom, L. J. Phys. Chem. 1996, 100, 16502.
J. Org. Chem, Vol. 69, No. 16, 2004 5239
